Verify Onchain Action
An independent signed safety verdict on a transfer, swap, approval or contract call before your agent signs it: catches drainers, honeypots, address poisoning and mandate breaches.
Free
Works with the AI Agents you already use
verify-onchain-action
Example session with this skill installed
My agent prepared approve(spender, uint256_max) for USDC on Base to swap 50 USDC on Uniswap. Check it before I sign.
- Read your context and instructions
- Compiled the verify-onchain-action
Sent the approval to /review as an onchain_action with your 500 USDC mandate. Verdict: approve_with_concerns. The unlimited allowance is far more than the 50 USDC swap needs, so approve exactly 50 USDC instead, and confirm the spender is Uniswap's canonical router on Base. Signed proof attached under ivv_proof.
Connects securely to your tools. The creator never sees your data.
About this skill
A wallet/onchain or trading MCP (like Base MCP, or Coinbase MCP with its orders_create /
orders_preview / convert_execute / transfer tools) lets you prepare a transaction — a trade,
transfer, swap, approval, bridge, or contract call. It does not tell you whether that action is
safe to sign / safe to place. Once signed, an on-chain action is irreversible, and the chain will happily execute a
sound-looking transaction that drains your wallet.
The thing you can't self-serve is judgment on your own proposed action. So before the irreversible step — before you sign, or before you ask a human to approve — get an independent verdict.
Before approving an on-chain action you can't fully vouch for, get a signed verdict. Attach the proof to whatever you hand a human or another agent.
This is NOT a replacement for the chain's own guarantees (a confirmed tx is already trustlessly recorded — you don't need anyone to verify that). It is the pre-approval check on the transaction's intent and risk — the part no blockchain checks for you.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.
The reviewer has no live chain data: it reasons from what you send and tells you what to confirm (for example, whether a spender is the canonical router). It does not sign or send transactions. Each verdict after the free calls is a paid /review call.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 12 days ago
- Free with an Agensi account
- Accounts
Babyblueviper
- Software
- curloranother HTTP client
- Optional
- invinoveritas API keyorUSDC on Base
Only /review needs the key (free to register, 3 free try-it calls) or USDC on Base via x402; verify-proof needs nothing.
Permissions
Allowed Hosts
Sends the proposed transaction description you choose to submit to api.babyblueviper.com. Never send private keys or seed phrases; the skill never needs them.